US9232217B2 - Method and apparatus for objective video quality assessment based on continuous estimates of packet loss visibility - Google Patents

Method and apparatus for objective video quality assessment based on continuous estimates of packet loss visibility Download PDF

Info

Publication number
US9232217B2
US9232217B2 US13/991,907 US201113991907A US9232217B2 US 9232217 B2 US9232217 B2 US 9232217B2 US 201113991907 A US201113991907 A US 201113991907A US 9232217 B2 US9232217 B2 US 9232217B2
Authority
US
United States
Prior art keywords
packet loss
bit stream
visibility
features
feature
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active, expires
Application number
US13/991,907
Other languages
English (en)
Other versions
US20130271668A1 (en
Inventor
Savvas Argyropoulos
Bernhard Feiten
Marie-Neige Garcia
Peter List
Alexander Raake
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Deutsche Telekom AG
Original Assignee
Deutsche Telekom AG
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Deutsche Telekom AG filed Critical Deutsche Telekom AG
Assigned to DEUTSCHE TELEKOM AG reassignment DEUTSCHE TELEKOM AG ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: RAAKE, ALEXANDER, Argyropoulos, Savvas, FEITEN, BERNHARD, GARCIA, MARIE-NEIGE, List, Peter
Publication of US20130271668A1 publication Critical patent/US20130271668A1/en
Application granted granted Critical
Publication of US9232217B2 publication Critical patent/US9232217B2/en
Active legal-status Critical Current
Adjusted expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N17/00—Diagnosis, testing or measuring for television systems or their details
    • H04N17/004—Diagnosis, testing or measuring for television systems or their details for digital television systems
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N17/00—Diagnosis, testing or measuring for television systems or their details
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/44—Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/85—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using pre-processing or post-processing specially adapted for video compression
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/85—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using pre-processing or post-processing specially adapted for video compression
    • H04N19/89—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using pre-processing or post-processing specially adapted for video compression involving methods or arrangements for detection of transmission errors at the decoder

Definitions

  • the present invention relates to a method and apparatus for video quality assessment based on the analysis of the visibility of packet losses that may occur during the transmission of a video sequence through an error-prone network.
  • QoS Quality of Service
  • QoE Quality of Experience
  • Quality estimation methods commonly support a distinguished estimation of the quality related to the coding (compression, Q cod ) of the video signal and the quality due to packet loss during transmission (Q trans ). It is an inherent property of IP-networks that (mostly due to temporal overload at some point in the network) IP packets may be lost. Some of these losses may be almost invisible to the customer while others may cause severe degradation of the video quality. Even if countermeasures against these losses are part of an IPTV distribution system, these countermeasures can never guarantee an absolute remedy. For instance, a retransmission request may take too long, or the retransmitted packet itself might get lost. Therefore there is always a nonzero probability, that fragmentary bit streams are transmitted to the end user device. These in turn can cause visible or audible degradations in the reconstructed video. Measurement categories may therefore also include values to express the probability for losses. Such values may include the expression of a “packet loss rate” and the “burstiness of loss events”.
  • the present invention provides a method for assessing the quality of a transmitted video signal sequence at a receiver side.
  • the method includes: a) capturing the input video bit stream and supplying it to a video bit stream analyzer; b) extracting at least one feature or a set of features from the captured input video bit stream by the bit stream analyzer; c) supplying the at least one extracted feature or set of features to a packet loss visibility estimation module; d) determining, by the packet loss visibility estimation module employing the supplied extracted features of the video bit stream, the continuous probability of visibility for each packet loss event which occurred within a specific time interval; and e) employing the continuous probability of packet loss visibility, determined by the packet loss visibility estimation module, as a weighting factor of the at least one feature or set of features extracted from the video bit stream to calculate an estimate of the overall quality, Q, of the transmitted video sequence.
  • Step (d) employs at least one bit stream feature selected from the group consisting of: frame type, average magnitude of motion vectors (AvgMv), average motion vector difference (AvgMvDiff), energy of residuals (ResEnergy), maximum number of partitions (MaxPartNr), number of non-decodable macroblocks (LostMbs), motion vector information (mv), and type of macroblocks (mb type).
  • Step (e) combines the packet loss visibility estimate (V) with the determined magnitude of distortion (EstErr) and the computed total number of impaired pixels due to packet loss (ErrProp).
  • FIG. 1 shows the architecture of the proposed video quality assessment model
  • FIG. 2 shows how frames are organized in a “group-of-pictures” (GOP) for video encoding
  • FIG. 3 shows the partitioning of macroblocks according to the H.264/AVC standard and the derivation of the transform coefficients after the application of the 4 ⁇ 4 integer transform
  • FIG. 4 shows the error propagation maps for four consecutive frames in the video sequence
  • FIG. 5 depicts a block diagram of a second embodiment of the present invention for the assessment of video quality
  • FIG. 6 depicts an example for the estimation of the induced visible distortion due to packet loss(es) in the frames of a video sequence.
  • Embodiments of the present invention provide a method and apparatus for the objective quality assessment of a video sequence.
  • the method and apparatus provide for the objective quality assessment of a video sequence based on at least one or a set of features extracted from the video bit stream and the prediction of continuous probabilistic estimates for the visibility of packet losses that may occur due to transmission of the video sequence over an error-prone network.
  • the method is a no-reference bit stream-based video quality assessment method, since it extracts information only from the received bit stream without the need for an explicit reference sequence.
  • an embodiment of the invention provides a method for assessing the quality of a transmitted video signal sequence at the receiver side, the method comprising the steps of:
  • the feature extraction according to step (b) may be performed directly by partly decoding the video bit stream.
  • the feature extraction according to step (b) is performed by full decoding of the video bit stream and by combining information from the reconstructed pixels of the video signal.
  • step (e) additionally weights the extracted feature(s) to reflect that each packet loss produces unequally perceived degradations to the transmitted video signal sequence.
  • the at least one feature or set of features is selected from the group comprising: frame type, average magnitude of motion vectors (AvgMv), average motion vector difference (AvgMvDiff), energy of residuals (ResEnergy), maximum number of partitions (MaxPartNr), number of non-decodable macroblocks (LostMbs), motion vector information (mv), and type of macroblocks (mb type). According to a further preferred embodiment, a combination of all these features is used.
  • the number of frames (ErrDur) degraded by a packet loss is determined.
  • the method further encompasses that a binary error propagation map is generated for each picture of the video signal sequence based on the motion vector information and the type of macroblocks.
  • the number of frames (ErrDur) degraded by a packet loss may be determined.
  • the total number of impaired pixels due to packet loss (ErrProp) is computed.
  • magnitude of distortion (EstErr) induced due to packet loss and error propagation is determined based on the motion vectors, the macroblock types, and the residuals.
  • the step (d) may be performed using a support vector regression technique by mapping the input feature vector to a high-dimensional feature space using a non-linear mapping function, and constructing a linear model in the feature space.
  • step (e) may combine the packet loss visibility estimate (V) with the determined magnitude of distortion (EstErr) and the computed total number of impaired pixels due to packet loss (ErrProp).
  • the method may further comprise between steps d) and e) the step of estimating the overall visible distortion generated by temporal pooling of all packet losses occurring within the video signal sequence.
  • an embodiment of the invention provides an apparatus for assessing the quality of a transmitted video signal sequence at the receiver side, the apparatus comprising:
  • a video bit stream analyzer receiving the captured input video bit stream, and configured for extracting at least one feature or a set of features from the captured input video bit stream
  • a packet loss visibility estimation module receiving the extracted feature or set of features, the packet loss visibility estimation module configured for determining the visibility of a packet loss occurred during transmission of the video signal by assigning a continuous estimate for each packet loss event occurred within a specific time interval;
  • a combiner for combining the packet loss visibility estimate determined by the estimation module with the extracted at least one feature or set of features to assess the overall quality, Q, of the transmitted video signal sequence.
  • the video bit stream analyzer is configured to perform the feature extraction directly by partly decoding the video bit stream.
  • the video bit stream analyzer is configured to perform the feature extraction by full decoding of the video bit stream and by combining information from the reconstructed pixels of the video signal.
  • the combiner is configured to additionally weight the extracted feature(s) to reflect that each packet loss produces unequally perceived degradations to the transmitted video signal sequence.
  • the at least one feature or set of features is selected from the group comprising: frame type, average magnitude of motion vectors (AvgMv), average motion vector difference (AvgMvDiff), energy of residuals (ResEnergy), maximum number of partitions (MaxPartNr), number of non-decodable macroblocks (LostMbs), motion vector information (mv), and type of macroblocks (mb type). According to a further preferred embodiment, a combination of all these features is used.
  • the apparatus is configured to determine, based on the extracted feature “frame type”, the number of frames (ErrDur) degraded by a packet loss.
  • the apparatus comprises an error propagation estimator being configured to generate a binary error propagation map for each picture of the video signal sequence based on the motion vector information and the type of macroblocks.
  • the number of frames (ErrDur) degraded by a packet loss may be determined.
  • the error propagation estimator is further configured to compute, based on the binary error propagation map, the total number of impaired pixels due to packet loss (ErrProp).
  • the apparatus comprises a distortion estimator being configured to determine the magnitude of distortion (EstErr) induced due to packet loss and error propagation based on the extracted motion vectors, the extracted macroblock types, and the extracted residuals.
  • a distortion estimator being configured to determine the magnitude of distortion (EstErr) induced due to packet loss and error propagation based on the extracted motion vectors, the extracted macroblock types, and the extracted residuals.
  • the packet loss visibility estimation module may be further configured to determine the visibility of the packet loss using a support vector regression technique by mapping the input feature vector to a high-dimensional feature space using a non-linear mapping function, and constructing a linear model in the feature space.
  • the combiner may further be configured to combine the packet loss visibility estimate (V) with the determined magnitude of distortion (EstErr) and the computed total number of impaired pixels due to packet loss (ErrProp).
  • the apparatus may further be configured to estimate the overall visible distortion generated by temporal pooling of all packet losses occurring within the video signal sequence prior to combining the packet loss visibility estimate determined by the estimation module with the extracted at least one feature or set of features.
  • the method and apparatus according to embodiments of the invention differentiate from the approaches proposed in the literature in at least three main points.
  • the method is different from the parametric models ([1, 2]) since it exploits the information from the bitstream to determine the objective video quality and is not based on packet header information only.
  • the extracted features from the bit stream enable the consideration of content dependency, the adaptation of the video quality measure to the spatio-temporal characteristics of the video sequences and take advantage of the properties of the human visual system (HVS) which are neglected in the parametric models.
  • HVS human visual system
  • the packet loss visibility of each packet loss is not estimated in a binary fashion (visible or invisible) but rather with a continuous value which denotes how probable is the packet loss under investigation to result in a perceivable and detectable degradation of the video sequence.
  • the granularity of the proposed method enables more accurate prediction of the visibility of packet losses due to network impairments.
  • all previously presented methods in the literature provide binary outputs for packet loss visibility estimation.
  • the output of the packet loss visibility classifier module is incorporated into the objective video quality assessment model to assess the quality of the bit stream. This method for video quality assessment explicitly exploits packet loss visibility estimation with continuous estimates for objective video quality assessment.
  • FIG. 1 depicts the block diagram of a preferred method and apparatus for the objective video quality assessment based on the estimation of the visibility of packet loss events.
  • it is shown that it is a no-reference bit-stream based method which extracts or computes eight features from the received bit-stream to evaluate the perceptual impact of every packet loss. Subsequently, the extracted features and the predicted value for the visibility of the packet loss are used to assess the overall quality of the video bit stream.
  • a probe device captures the bit stream and extracts or computes a number of features that are utilized for the video quality assessment.
  • the features are either extracted directly by partly decoding the bit stream or combining information from the reconstructed pixels (full decoding of the bit stream).
  • the extracted features are fed to the module which is responsible for determining the visibility of each packet loss event.
  • This module assigns a continuous estimate of visibility for each packet loss event (isolated packet losses or combination of events) that occurs within a specific time duration.
  • the probability estimate of the packet loss visibility module is combined with the previously extracted features to assess the overall quality of the sequence. Specifically, based on the probability estimates of the visibility of each packet loss, the extracted features are weighted in the final algorithm to reflect that every packet loss produces unequally perceived degradations to the video sequence.
  • the probe device outputs the predicted value of video quality, Q.
  • the following describes the features preferably extracted from the bit stream to determine the visibility of packet losses and the objective quality of the video sequences.
  • the analysis is based on streams encoded according to the H.264/AVC video coding standard (T. Wiegand, G. J. Sullivan, G. Bjontegaard, and A. Luthra, “Overview of the H.264/AVC video coding standard,” IEEE Trans. on Circuits and Systems for Video Technology, vol. 13, no. 7, July 2003), but the method could be applied to any video coding standard with minor modifications for the computation of the corresponding features.
  • One of the properties that need to be computed for the estimation of the perceptual degradation caused by a packet loss is the “frame type” of all frames and in particular the frame affected by the loss.
  • the possible values for the “frame type” property include “Intra-frame” or “Key-Frame” (below called I-frame), “Predicted-Frame” (below called P-frame) and “Bidirectional-Frame” (below called B-frame).
  • I-frames can be decoded without referencing information of any prior frames.
  • P-frames depend on one or more predecessors called “reference frames”, because the information transmitted for a P-frame mainly consists of the difference between the video-frame it describes and its references.
  • the chain of video frames between two successive I-frames is called “Group of Pictures” (GoP).
  • GoP Group of Pictures
  • P and B-frames in a GoP follow a more or less strict pattern like the typical GoP-pattern: “I, B, B, P, B, B, P . . . ”.
  • N denote the GOP-length, that is the distance between two successive I-frames.
  • the first packet loss occurs t frames after the first I-frame of a specific GOP
  • the first feature that is extracted is the number of frames that are degraded due to the packet loss, called ErrDur. This property is computed as (see FIG. 2 ):
  • the average magnitude of motion vectors is preferably computed which reflects the mean motion vector in the horizontal and vertical directions of all macroblocks (mb) affected by the loss.
  • v n [v n,x , v n,y ] denote the motion vector of the n-th macroblock in the frame in which the packet loss occurred.
  • AvgMv is computed as:
  • L is the cardinality of the set of the macroblocks that are lost in frame n due to the packet loss.
  • the average motion vector difference may be computed from the bit stream.
  • AvgMvDiff may be computed from the bit stream.
  • H.264/AVC to exploit the redundancy among motion vectors, only the difference between the motion vector of a block and its predicted motion vector from neighboring macroblocks is included in the bit stream. This feature is extremely meaningful in sequences with predictable movement (e.g., panning), where the average motion may be large, but the motion vector differences encoded in the bit stream are small.
  • vd n [vd n,x , vd n,y ] denote the motion vector difference of the n-th macroblock in the frame in which the packet loss occurred.
  • AvgMvDiff is computed as:
  • L is the cardinality of the set of the macroblocks that are lost in frame n due to the packet loss.
  • this information is estimated from their temporal neighbours. That is, for every missing macroblock, the motion vector information is recovered from the co-located macroblock in the previous correctly received frame.
  • ResEnergy Another feature that is preferably extracted from the received video bit stream is the energy of the residual, ResEnergy, which is contained in the missing macroblocks and is computed from the transform coefficients of the prediction error.
  • MaxPartNr the maximum number of partitions in the frame where the loss occurs.
  • each 16 ⁇ 16 macroblock can be further split in smaller blocks for the motion estimation process, i.e. blocks of size 16 ⁇ 8, 8 ⁇ 16, or 8 ⁇ 8. If the 8 ⁇ 8 partition size is chosen, each 8 ⁇ 8 partition may be further divided in sub-partitions of size 8 ⁇ 8, 8 ⁇ 4, 4 ⁇ 8, or 4 ⁇ 4 luma samples.
  • MaxPartNr is equal to the maximum number of partitions of the correctly received macroblocks in the frame where the loss occurs. If all the macroblocks are erased (that is, if the whole frame is lost) then the parameter is derived from the previous received frame.
  • the sixth feature that is extracted from the received bit stream is the number of macroblocks that are non-decodable due to the packet loss and need to be concealed, called LostMbs below.
  • a binary error propagation map may be generated for each picture, which denotes the pixels in each picture that are impacted by the packet loss(es) either due to the innovation error or due to the propagation of the error to the subsequent frames of the same GOP due to the predictive coding.
  • k 1, . . .
  • K is the number of frames that contain impaired pixels due to the packet loss.
  • the value for those pixels that are either impaired due to the packet loss or reference those areas and are likely to be erroneous is set to one, otherwise the value is set to zero.
  • I ⁇ ( x , y , k ) ⁇ 1 , if ⁇ ⁇ pixel ⁇ ⁇ in ⁇ ⁇ position ⁇ ⁇ ( x , y ) ⁇ ⁇ of ⁇ ⁇ frame ⁇ ⁇ k ⁇ ⁇ is ⁇ ⁇ impaired 0 , if ⁇ ⁇ pixel ⁇ ⁇ in ⁇ ⁇ position ⁇ ⁇ ( x , y ) ⁇ ⁇ of ⁇ ⁇ frame ⁇ ⁇ k ⁇ ⁇ is ⁇ ⁇ not ⁇ ⁇ impaired Eq . ⁇ ( 5 )
  • the value in the propagation map array is set to one for all pixels that are not identical between the original video sequence and the reconstructed image of the decoder within the probe due to the error caused by the packet loss and its propagation.
  • An example of the derivation of the error propagation maps is depicted in FIG. 4 .
  • FIG. 4 depicts the error propagation maps for four consecutive frames in the video sequence to demonstrate how the initial error caused by a packet loss spreads into the following frames that reference the initial frame for their prediction.
  • the dark region corresponds to the area that cannot be decoded because the information for these pixels was contained in a packet which was erased.
  • the decoder needs to conceal (usually using temporal or spatial neighbors) this area to replace the missing pixels. Since the concealment can not recover the original information exactly, some degradation is introduced to that frames.
  • FIGS. 4 b , 4 c , and 4 d show how this error propagates into the following frames. Dark areas correspond to pixels that reference information from the initially lost (and concealed) region and are thus prone to be erroneous as well.
  • the packet loss occurs in the first frame (the dark area in the figure denotes the macroblocks that have been erased due to the packet loss) and propagates to the subsequent frames that reference the affected area of the first frame for the prediction process. It is noted that in alternative embodiments the error propagation map may be estimated at the macroblock level, meaning that propagation is tracked for every macroblock instead for every pixel.
  • ErrProp the total number of impaired pixels due to a packet loss
  • EstErr another feature may be extracted based on the motion vectors, the macroblock types and the residuals, termed EstErr, to quantify the magnitude of the distortion induced due to the packet loss and due to error propagation which enables the estimation of the induced mean squared error (MSE) in a no-reference manner.
  • MSE mean squared error
  • the method for estimating the induced distortion is out of scope of this invention, thus any method known to the skilled person could be used, for example the method proposed in M. Naccari, M. Tagliasacchi, and S. Tubaro, “No-reference video quality monitoring for H.264/AVC coded video,” IEEE Trans. on Multimedia, vol. 11, no. 5, pp. 932-946, August 2009.
  • the aforementioned eight features are extracted for every packet loss that occurs in a GOP.
  • the maximum value is selected for the ErrorDur, and MaxPartNr, while for the other features their sum is considered as the final features.
  • the features extracted from every packet loss within a GOP can be combined differently, for example by computing the mean, the median, or any other function.
  • SVR Support Vector Regression
  • b which are used during the evaluation stage.
  • the value of b is selected to be equal to 1.27. However, any other value could be used.
  • the algorithm receives an input feature vector f (as described in Eq. (7)) and the predicted value of the visibility of each loss, V, is given by:
  • K(•,•) is the kernel function
  • the kernel function which is also known as the Radial Basis Function (RBF):
  • the value of y is selected to be equal to 2. Also, any other kernel function could be used in other embodiments.
  • the predicted value V for the visibility of a packet loss can be converted into a binary value, if it is desirable to classify a packet loss into one of the following two cases: visible or invisible. In that case, the binary predicted value of a packet loss, denoted as V B , can be computed as:
  • V B ⁇ 0 , if ⁇ ⁇ V ⁇ T 1 , if ⁇ ⁇ V ⁇ T Eq . ⁇ ( 10 )
  • T is a threshold value to classify the continuous estimate of packet loss visibility in a binary value.
  • the suggested value is 0.7, but any other value could be used.
  • the stored parameters are used and the visibility V of each packet loss event is predicted as defined in Eq. (8).
  • the extracted features that were mentioned above are refined so that they reflect the predicted value of the visibility of each packet loss and are weighted accordingly, as analyzed in the following section.
  • the proposed bit stream based video quality assessment model is a combination of the degradation caused by the compression of the video sequences, Q cod , and the quality degradation due to packet loss during transmission (Q trans ).
  • the index i is used to denote a particular single packet loss event that occurs during the sequence and ⁇ (•,•) represents any suitable function.
  • the overall degradation due to transmission errors is a function of the visibility of every packet loss, the error that is induced to the frame where the packet loss occurs and the propagation of the error into the subsequent frames.
  • NrImpPx k i denote the number of impaired pixels in frame k caused by the i-th packet loss, or equivalently
  • H, W are the height and width of each frame of the video sequence, respectively.
  • VisEstErr k i VisEstErr 0 i ⁇ NrImpPx k i NrImpPx 0 i Eq . ⁇ ( 17 )
  • the dark area in frame with index t corresponds to the term NrImpPx 0 i
  • the dark area in the subsequent frames indeces t+1, . . . , t+m
  • VisEstErr GOP ⁇ i ⁇ VisEstErr i Eq . ⁇ ( 19 )

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • General Health & Medical Sciences (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)
  • Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
US13/991,907 2010-12-10 2011-09-02 Method and apparatus for objective video quality assessment based on continuous estimates of packet loss visibility Active 2032-08-22 US9232217B2 (en)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
EP10194580 2010-12-10
EP10194580 2010-12-10
EP10194580.6 2010-12-10
PCT/EP2011/065192 WO2012076203A1 (en) 2010-12-10 2011-09-02 Method and apparatus for objective video quality assessment based on continuous estimates of packet loss visibility

Publications (2)

Publication Number Publication Date
US20130271668A1 US20130271668A1 (en) 2013-10-17
US9232217B2 true US9232217B2 (en) 2016-01-05

Family

ID=44719874

Family Applications (1)

Application Number Title Priority Date Filing Date
US13/991,907 Active 2032-08-22 US9232217B2 (en) 2010-12-10 2011-09-02 Method and apparatus for objective video quality assessment based on continuous estimates of packet loss visibility

Country Status (11)

Country Link
US (1) US9232217B2 (ko)
EP (1) EP2649801B1 (ko)
JP (1) JP6022473B2 (ko)
KR (1) KR101857842B1 (ko)
CN (1) CN103283239B (ko)
BR (1) BR112013014367B1 (ko)
ES (1) ES2541282T3 (ko)
HR (1) HRP20150481T1 (ko)
HU (1) HUE026522T2 (ko)
PL (1) PL2649801T3 (ko)
WO (1) WO2012076203A1 (ko)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150341667A1 (en) * 2012-12-21 2015-11-26 Thomson Licensing Video quality model, method for training a video quality model, and method for determining video quality using a video quality model
US20150365662A1 (en) * 2013-02-07 2015-12-17 Thomson Licensing Method And Apparatus For Context-Based Video Quality Assessment
US9716881B2 (en) * 2013-02-07 2017-07-25 Thomson Licensing Method and apparatus for context-based video quality assessment
US11216742B2 (en) 2019-03-04 2022-01-04 Iocurrents, Inc. Data compression and communication using machine learning

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2468878B (en) * 2009-03-24 2014-11-19 Strategy & Technology Ltd A device comprising at least one receiver for a testing device
US8804815B2 (en) * 2011-07-29 2014-08-12 Dialogic (Us) Inc. Support vector regression based video quality prediction
GB2497915B (en) * 2011-10-25 2015-09-09 Skype Estimating quality of a video signal
US9661348B2 (en) * 2012-03-29 2017-05-23 Intel Corporation Method and system for generating side information at a video encoder to differentiate packet data
US9769501B2 (en) 2012-07-17 2017-09-19 Thomson Licensing Video quality assessment at a bitstream level
WO2014032451A1 (en) * 2012-08-27 2014-03-06 Thomson Licensing Method and apparatus for estimating motion homogeneity for video quality assessment
GB2513090B (en) 2013-01-28 2019-12-11 Microsoft Technology Licensing Llc Conditional concealment of lost video data
CN103108210B (zh) * 2013-03-07 2015-04-15 福州大学 基于空域复杂度的无参考视频质量评价方法
US9674515B2 (en) * 2013-07-11 2017-06-06 Cisco Technology, Inc. Endpoint information for network VQM
CN104092960B (zh) * 2014-07-26 2017-06-06 福州大学 一种基于视觉主观感受的视频源切换方法
JP6499993B2 (ja) * 2016-05-18 2019-04-10 株式会社ソニー・インタラクティブエンタテインメント 情報処理装置、情報処理システム、および情報処理方法
CN111372071B (zh) * 2018-12-25 2022-07-19 浙江宇视科技有限公司 一种搜集视频图像异常信息的方法及装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006033722A (ja) 2004-07-21 2006-02-02 Nippon Telegr & Teleph Corp <Ntt> 映像品質管理方法および映像品質管理システム
JP2008301026A (ja) 2007-05-30 2008-12-11 Nippon Telegr & Teleph Corp <Ntt> 映像品質推定方法、装置、およびプログラム
EP2106153A1 (en) 2008-03-28 2009-09-30 Deutsche Telekom AG Audio-visual quality estimation
US20100110199A1 (en) 2008-11-03 2010-05-06 Stefan Winkler Measuring Video Quality Using Partial Decoding
US20130044224A1 (en) * 2010-04-30 2013-02-21 Thomson Licensing Method and apparatus for assessing quality of video stream

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1804519A4 (en) * 2004-10-18 2010-01-06 Nippon Telegraph & Telephone OBJECTIVE VIDEO QUALITY EVALUATION DEVICE, EVALUATION METHOD AND PROGRAM

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006033722A (ja) 2004-07-21 2006-02-02 Nippon Telegr & Teleph Corp <Ntt> 映像品質管理方法および映像品質管理システム
JP2008301026A (ja) 2007-05-30 2008-12-11 Nippon Telegr & Teleph Corp <Ntt> 映像品質推定方法、装置、およびプログラム
EP2106153A1 (en) 2008-03-28 2009-09-30 Deutsche Telekom AG Audio-visual quality estimation
US20100110199A1 (en) 2008-11-03 2010-05-06 Stefan Winkler Measuring Video Quality Using Partial Decoding
US20130044224A1 (en) * 2010-04-30 2013-02-21 Thomson Licensing Method and apparatus for assessing quality of video stream

Non-Patent Citations (14)

* Cited by examiner, † Cited by third party
Title
Fuzheng Yang et al: "No-Reference Quality Assessment for Networked Video via Primary Analysis of Bit Stream", IEEE Transactions on circuits and systems for video technology, IEEE service center, Piscataway, NJ, US, vol. 20, No. 11, Nov. 1, 2010, pp. 1544-1554, XP011319799, ISSN: 1051-8215.
K. Yamagishi and T. Hayashi, "Parametric Packet-Layer Model for Monitoring Video Quality of IPTV Services," in Proc. of IEEE Int. Conf. on Communications, Dec. 2008.
M. Garcia and A. Raake, "Parametric packet-layer video quality model for IPTV," Int. Conf. on Information Science, Signal Processing and their Applications (ISSPA), Kuala-Lumpur, May 2010.
M. Naccari, M. Tagliasacchi, and S. Tubaro, "No-reference video quality monitoring for H.264/AVC coded video," IEEE Trans. on Multimedia, vol. 11, No. 5, pp. 932-946, Aug. 2009.
Mu Mu et al: "A Discrete perceptual impact evaluation quality assessment framework for IPTV services," Multimedia and expo (ICME), 2010 IEEE International conference on, IEEE, Piscataway, NJ, USA, Jul. 19, 2010, pp. 1505-1510, XP031761609, ISBN: 978-1-4244-7491-2.
N. Staelens et al., "Viqid: A no-reference bit stream-based visual quality impairment detector," Proceedings of the 2010 second international workshop on quality of multimedia experience (QOMEX 2010), Trondheim, Norway, IEEE, Piscataway, NJ, USA, Jun. 21, 2010, pp. 206-211, XP031712774.
S. Kanumuri, P. C. Cosman, A. R. Reibman, V. A. Vaishampayan, "Modeling packet-loss visibility in MPEG-2 video," IEEE Trans. On Mulitmedia, vol. 8, No. 2, Apr. 2004, pp. 341-355.
S. Kanumuri, S. B. Subramanian, P. C. Cosman, A. R. Reibman, "Predicting H.264 packet loss visibility using a generalized linear model," Image Processing, 2006 IEEE International conference on, IEEE, PI, Oct. 1, 2006, pp. 2245-2248, XP031049119, ISBN: 978-1-4244-0480-3.
Savvas Argyropoulos et al: "No-reference bit stream model for video quality assessment of h.264/AVC video based on packet loss visibility," acoustic, speech and signal processing (ICASSP), 2011 IEEE international conference on, IEEE, May 22, 2011, pp. 1169-1172, XP032000951, DOI: 10.1109/ICASSP.2011.5946617, ISBN: 978-1-4577-0538-0.
T. Wiegand, G. J. Sullivan, G. Bjontegaard, and A. Luthra, "Overview of the H.264/AVC video coding standard," IEEE Trans. on Circuits and Systems for Video Technology, vol. 13, No. 7, Jul. 2003.
T.L. Lin, S. Kanumuri, Y. Zhi, D. Poole, P.C. Cosman, and A.R. Reibman, ,,A versatile model for packet loss visibility and its application to packet prioritization, IEEE Trans. On Image Processing, vol. 19, No. 3, pp. 722-735, Mar. 2010.
Takahashi A. et al: "Standardization activities in the ITU for a QoE assessment of IPTV," IEEE Communications magazine, IEEE Service center, Piscataway, US, vol. 46, No. 2, Feb. 1, 2008, pp. 78-84, XP011206259, ISSN: 0163-6804.
Winkler S et al: "The Evolution of Video Quality Measurement: From PSNR to Hybrid Metrics," IEEE Transaction on broadcasting, IEEE service center, Piscataway, NJ, US, vol. 54, No. 3, Sep. 1, 2008, pp. 660-668, XP011229276, ISSN: 0018-9316, DOI: 10.1109/TBC.2008.2000733.
Y. J. Liang, J. G. Apostolopoulos, and B. Girod, "Analysis of packet loss for compressed video: effect of burst losses and correlation between error frames," IEEE Transactions on Circuits and Systems for Video Technology, IEEE Service Center, Piscataway, NJ, US, vol. 18, No. 7, Jul. 1, 2008, pp. 861-874, XP011226407, ISSN: 1051-8215, DOI: 10.1109/TCSVT.2008.923139.

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150341667A1 (en) * 2012-12-21 2015-11-26 Thomson Licensing Video quality model, method for training a video quality model, and method for determining video quality using a video quality model
US20150365662A1 (en) * 2013-02-07 2015-12-17 Thomson Licensing Method And Apparatus For Context-Based Video Quality Assessment
US9716881B2 (en) * 2013-02-07 2017-07-25 Thomson Licensing Method and apparatus for context-based video quality assessment
US9723301B2 (en) * 2013-02-07 2017-08-01 Thomson Licensing Method and apparatus for context-based video quality assessment
US11216742B2 (en) 2019-03-04 2022-01-04 Iocurrents, Inc. Data compression and communication using machine learning
US11468355B2 (en) 2019-03-04 2022-10-11 Iocurrents, Inc. Data compression and communication using machine learning

Also Published As

Publication number Publication date
BR112013014367B1 (pt) 2022-07-12
ES2541282T3 (es) 2015-07-17
US20130271668A1 (en) 2013-10-17
CN103283239A (zh) 2013-09-04
CN103283239B (zh) 2016-10-05
BR112013014367A2 (pt) 2021-07-27
KR20140008508A (ko) 2014-01-21
WO2012076203A1 (en) 2012-06-14
JP6022473B2 (ja) 2016-11-09
PL2649801T3 (pl) 2015-08-31
EP2649801B1 (en) 2015-04-15
HUE026522T2 (en) 2016-06-28
EP2649801A1 (en) 2013-10-16
HRP20150481T1 (hr) 2015-06-05
KR101857842B1 (ko) 2018-05-14
JP2014502113A (ja) 2014-01-23

Similar Documents

Publication Publication Date Title
US9232217B2 (en) Method and apparatus for objective video quality assessment based on continuous estimates of packet loss visibility
US9131216B2 (en) Methods and apparatuses for temporal synchronisation between the video bit stream and the output video sequence
US9025673B2 (en) Temporal quality metric for video coding
US10075710B2 (en) Video quality measurement
KR101834031B1 (ko) 비디오 신호의 인코딩 및 전송 동안에 비디오 신호의 품질 평가 방법 및 장치
US9077972B2 (en) Method and apparatus for assessing the quality of a video signal during encoding or compressing of the video signal
Argyropoulos et al. No-reference bit stream model for video quality assessment of H. 264/AVC video based on packet loss visibility
Reibman et al. Characterizing packet-loss impairments in compressed video
Chen et al. Hybrid distortion ranking tuned bitstream-layer video quality assessment
Yamada et al. No-reference video quality estimation based on error-concealment effectiveness
Reibman et al. Predicting packet-loss visibility using scene characteristics
CN104396238B (zh) 比特流级别的视频质量评估
US9723266B1 (en) Lightweight content aware bit stream video quality monitoring service
Seeling et al. Offset distortion traces for trace-based evaluation of video quality after network transport
Liu et al. Perceptual quality measurement of video frames affected by both packet losses and coding artifacts
Farrugia et al. Objective video quality metrics for HDTV services: A survey
Yang et al. Temporal quality evaluation for enhancing compressed video
Seeling et al. Offset distortion traces for trace-based evaluation of video quality after network transport (extended version)
Naccari et al. Subjective evaluation of a no-reference video quality monitoring algorithm for H. 264/AVC video over a noisy channel
US20140119460A1 (en) Method and device for assessing packet defect caused degradation in packet coded video

Legal Events

Date Code Title Description
AS Assignment

Owner name: DEUTSCHE TELEKOM AG, GERMANY

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:ARGYROPOULOS, SAVVAS;FEITEN, BERNHARD;GARCIA, MARIE-NEIGE;AND OTHERS;SIGNING DATES FROM 20130524 TO 20130529;REEL/FRAME:030741/0298

FEPP Fee payment procedure

Free format text: PAYOR NUMBER ASSIGNED (ORIGINAL EVENT CODE: ASPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

STCF Information on status: patent grant

Free format text: PATENTED CASE

MAFP Maintenance fee payment

Free format text: PAYMENT OF MAINTENANCE FEE, 4TH YEAR, LARGE ENTITY (ORIGINAL EVENT CODE: M1551); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

Year of fee payment: 4

MAFP Maintenance fee payment

Free format text: PAYMENT OF MAINTENANCE FEE, 8TH YEAR, LARGE ENTITY (ORIGINAL EVENT CODE: M1552); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

Year of fee payment: 8